Signal2026-07-01
arXiv

Reinforcement Learning with Metacognitive Feedback Elicits Faithful Uncertainty Expression in LLMs

Open primary sourceRead on Substack